How strong is the relationship between the large-scale environment and tropical cyclone climatology in climate models?

Friday, December 14, 2018 - 17:15
Add to Calendar

We explore the relationship between the climatological characteristics of tropical cyclones (TCs) in climate models and the simulated large-scale environments within which they occur. If a model has a mean climate that in observations would be associated with more and/or more intense TCs, does that model actually have a TC climatology with those characteristics, compared to another model whose mean climate would in observations be associated with fewer and/or weaker TCs? We consider 30 climate models with a wide range of horizontal resolutions.

As expected, there is a clear relationship between horizontal resolution, with higher-resolution models producing stronger TCs. Nonetheless, models with the same horizontal resolution can have very different TC climatologies.

We explore whether there are systematic relationships across this large multi-model ensemble between TC statistics - such as climatological number of tropical cyclones (NTC) or accumulated cyclone energy (ACE) - and the climatological environmental fields usually associated with TC activity, such as potential intensity or vertical shear, or genesis indices. For low-resolution models, there is no association between environmental fields and TC activity metrics. As model resolution increases, this relationship is weak for some variables. Therefore, analyses of model-model differences in the large-scale model environmental fields are not informative in understanding the differences in TC climatology across models, even though differences in those same fields across time and space in observations can be highly informative in explaining observed variations in TC activity. Model-model differences in TC statistics instead result primarily from the different ways in which key model characteristics – such as model physics and dynamical core – determine the behavior of TCs within a fixed environment. These conclusions point to the need for more complex “process-oriented” diagnostics, such as the ones developed in Kim et al. 2018, in order to understand how model properties control simulated TC activity, and thus how to reduce model biases in TC activity.

Kim, D., and co-authors, 2018. Process-oriented diagnosis of tropical cyclones in high-resolution GCMs. J. Climate, 31, 1685-1702, doi: 10.1175/JCLI-D-17-0269.1.

Link for More Information: